Release Burndown chart - help to understand (in clear language or examples)

[Debrenee Fisher]

I have ready through the Jira documentation on how to understand the Release Burndown chart but am more confused.  The documentation uses one example but does not clarify how the assumptions were created.  Do any of you have a good handle on this and can give a simple example? 

I am trying to understand the predictive logic and what I can do to cleanup the data to be closer to what the team(s) believe it should be.